Dear All,

I would like some assistance in an issue we have had, around a week ago we
had a network failure, where our Active/Passvie Cluster failed over. The
orginal Active Node (node A) still tried keeping the Cluster IP address for 3
hours, although the resources failed over to Node B ok.

We think this is related to the virtual ip address not updating the ARP
cache on the routers, all our network routers rebooted, the spanning tree
took around 30 sec's to update but the ARP cache only updates every 3 hours
(i am told by our network team).

My question is that i have seen on
http://www.microsoft.com/technet/its...g/tcpip2k.mspx
you can change the number of ARP broadcasts, but is there a setting that
could delay the gratuitous Address Resolution Protocol (ARP) request when a
failover occurs?

Also when a failover occurs what is the time span that the ARP requests get
sent out, i.e once failed over does the new Active cluster send the requests
out all at once or wait for a time out then send the final 2 broadcasts.

All this would be helpful to try and stop the issue happening again.

Regards
Keith